This invention relates to a process for producing 2-methyleneglutaronitrile (hereinafter sometimes referred to as MGN) by liquid phase-dimerization of acrylonitrile (hereinafter sometimes referred to as AN). More particularly, this invention relates to a process for recovering MGN efficiently from the liquid dimerization product.
Hitherto, a process for preparation of MGN comprising contacting AN in liquid phase with a catalyst composed of a metal halide and a trialkylamine has been known. According to this process, an MGN solution is obtained, but the MGN cannot always be recovered readily from the resulting liquid reaction product.
Distillation methods may be used to recover MGN from the liquid reaction product. However, MGN itself is a readily polymerizable compound, and, moreover, the polymerization of MGN readily occurs when it is heated in the presence of a catalyst for the synthesis of MGN. Thus, the recovery of MGN by distillation is very low, and, moreover, formation of MGN polymers causes trouble in the operation of apparatus. Therefore, such a distillation method is not practical on a commercial scale.
It may be readily anticipated that the problems of the marked polymerization in the case of recovering MGN by distillation from the reaction liquid product can be solved to some extent by deactivating or removing the remaining catalyst. Accordingly, a method has been proposed which comprises deactivating the remaining catalyst by treating the reaction product with an aqueous acid solution or an aqueous alkali solution, transferring it to the aqueous layer, and then separating the aqueous layer by stationary separation. According to this method, however, separation of the aqueous layer from the oily layer is very difficult because an emulsion forms, and, moreover, the separation operation is very complicated because lumps of polymer form. In addition, the treatment of the drainage water containing a metal halide and a halogen is also required. Thus this method is very troublesome as a commercial process.